Webinar Recording · 4 August 2026

Admissions 2030: The Intelligent Front Door.

The full recording from our live session with Ben Rogers and Lesley O'Keeffe. A fireside chat on where admissions is actually heading by 2030, and what a realistic first step looks like.

The questions, organised by topic

From the live discussion and the audience Q&A, answered by Ben Rogers, Lesley O'Keeffe and Alistair Sergeant. Pick a topic, then click any question to read the answer.

Break your process down into steps and work out which ones AI could genuinely help with: grade verification, checking information across systems where you don't have a formal integration, and the back-and-forth engagement with applicants who need quick answers. Rules-based processes, like fee assessment, are a strong place to start too.Lesley O'Keeffe & Ben Rogers

Clip: "How can I actually use AI to enhance the admissions journey?"

Map your processes first. Before any technology goes on top of anything, understand your data and what it actually means, because poor data in means poor data out. Then talk to your team about the task that makes them "lose the will to live" through sheer repetition. Removing that one thing is often the fastest way to earn real buy-in.Ben Rogers & Lesley O'Keeffe

Clip: "What's one thing to do Monday morning to not be left behind?"

Niico has an ROI business-case calculator that maps standard admissions processes, so you can see where the real opportunity is before committing to anything. Niico is also running further sessions from September, with institutions including the University of Hull, ARU and KCL sharing how they've started.Alistair Sergeant
A lot of registry and enrolment pain shows up as paperwork, regulation, and administrative burden: bigger demands from students on getting information back, more documentation to process, more red tape. AI can take that burden off the person doing the work, making the difficult tasks far more automated and straightforward, so the system does the heavy lifting rather than the team.Ben Rogers
Yes. Niico will circulate a brief of the relevant products and tools following the session.Alistair Sergeant
Automation handles a defined, repeatable process with clear rules. AI comes in when the data isn't structured and needs judgement applied to it. Most of the confusion in the sector comes from people hearing "AI" and picturing ChatGPT, when it's a much broader toolkit than that.Lesley O'Keeffe
Most institutions are aiming for the top of the maturity map: full, joined-up, intelligent automation. In reality, based on Niico's own work and JISC/EY survey data, most institutions are still at the earliest stages. There's a long way to go before that top stage is realistic for most.Alistair Sergeant
The sector wants to change, but often doesn't know how. Much of the resistance comes down to language: a lack of understanding of what AI or agents actually do, versus what they're assumed to do. When people hear "AI", they think ChatGPT, when it's a much broader toolkit. There's a genuine piece of work to be done in upskilling and education before mindsets shift, and that has to happen alongside giving teams the headspace to engage with it, not on top of an already relentless workload.Lesley O'Keeffe

Clip: "Automation is the default, not the add-on" — how ready is the sector, strategy and culture, moving mindsets

The honest answer: it may reduce headcount at some institutions, and pretending otherwise given current mergers and cuts isn't credible. But the argument isn't "AI replaces people". It's freeing people from repetitive work so they can focus on the admissions decisions that genuinely need a human: the borderline cases, the difficult conversations, the ones that need real care.Ben Rogers & Lesley O'Keeffe

Clip: "Is AI going to take the jobs?" (Ben & Lesley, 28:04 to 31:21)

Yes. Lesley's own institution has a mitigating circumstances agent live today, checking a genuinely multi-system process. In the first few weeks it handled around 1,000 applications the team never had to touch manually. The advice: start small and tangible, build trust in the result, then scale, rather than trying to automate an entire process in one move.Lesley O'Keeffe

Clip: The mitigating circumstances agent (Lesley O'Keeffe, 34:20 to 35:48)

Three principles. Baseline before you start, two or three weeks of volume, response time and backlog data is enough and it costs nothing, and most ROI cases collapse because nobody measured the before. Measure the work, not the people, headcount is the wrong unit and the one everyone's afraid of. Our own reference point at ARU: 700 enquiry emails a day reduced to roughly 50 needing a human, a capacity number, not a redundancy number. Say the intent out loud on day one, whether that's holding headcount through rising volume or genuinely reducing roles, ambiguity is what people fill with fear, and staff will work it out either way. One thing that buys real credibility: publish the failure modes alongside the wins, escalation rate, what the system got wrong, what humans had to fix, teams trust the numbers far more when you show the misses too. And be honest about the real comparison: for most institutions it isn't AI versus the current team, it's AI versus the same team absorbing 20% more volume with no budget for more staff.Alistair Sergeant
Money is tight almost everywhere, and most institutions carry a legacy of bruising, badly-run change projects, so waiting for your existing provider is understandable. The drawback is speed: providers build for the average case, not for what makes your institution different, so real innovation gets lost in the wait. Starting a procurement process today likely means a 2030 go-live at best, by which point the technology has moved on again. The alternative the paper argues for: incremental automation layered onto what you already run, rather than one large-scale replacement project.Lesley O'Keeffe & Ben Rogers

Clip: "Should we invest in a new SRS, or wait?"

Yes. Niico works with institutions from a few hundred students up to several thousand, and the technology is built to scale to that size. The starting point is the same regardless of institution size: find one process worth automating that stacks up on time saved and ROI. Being smaller does not put this out of reach.Alistair Sergeant
It depends on where admissions sits at your institution, but realistically it needs all three working in partnership. IT support is non-negotiable for any successful implementation, but the drive has to come from the business side too, or projects stall on anxiety rather than momentum. Executive-level backing matters for smoothing over the inevitable bumps along the way.Ben Rogers
That accountability doesn't go away, and it shouldn't. The processes worth automating first are the well-defined, high-volume ones where a human stays firmly in charge of anything genuinely borderline or difficult. The agent does the heavy lifting; the accountable person still makes the call that matters.Ben Rogers & Lesley O'Keeffe
The goal isn't to avoid dependency, every institution is already dependent on its SRS, its CRM, its finance system. The goal is to make that dependency reversible. Own three things regardless of supplier: your data, exportable on demand in open formats, without a professional-services charge to get it out; your process logic, documented outside the vendor's configuration screen so it could be rebuilt elsewhere; and your team's capability, if nobody can still do the work manually, you haven't automated a process, you've forgotten one. Keep AI as a layer sitting on top of your systems of record rather than becoming the record itself, if the layer goes, the record still stands. Ask any supplier two questions: are you locked to a single model provider, or can you swap it without a full re-implementation, and what are the exit terms and the cap on annual price escalation? Then apply the degradation test: if this disappeared the week before Clearing, is the honest answer "slower and more expensive", or "we stop"? If it's the second, you've over-integrated. Start with work where failure is survivable, enquiry handling and document checks, not offer decisions.Alistair Sergeant
They're not necessarily in competition. For generic information, it may not matter to an applicant whether they're speaking to a bot or a person. The more interesting opportunity sits at the deep-personalisation end: AI that spots when a piece of research or news matches an applicant's stated interests, and surfaces that automatically, may actually go further than peer-to-peer communication can on its own.Lesley O'Keeffe
Yes, explicitly. The paper isn't arguing for removing face-to-face conversations for the decisions that need real judgement. What changes is everything around them: the chasing, the document checks, the repeat questions, freeing up time and headspace for the conversations that actually need a person.Ben Rogers

Clip: "Do borderline decisions still need a human?" (Utopia vs Reality of Admissions 2030)

Quality rarely degrades from volume itself, it degrades for two separate reasons, and they need separate answers. Mismatch: growth harms quality when you recruit students who were never going to succeed on that course. The fix is measurement discipline, track continuation and completion by recruitment channel and entry route, not just applications and offers, a channel that converts brilliantly and continues badly is costing you money and outcomes at the same time. Capacity strain: growth plans should be stress-tested against staff-to-student ratios, timetable and estate, placement availability, assessment load and student support demand, not just admissions throughput, and where the honest limit is reached, cap the intake. In England, the B3 thresholds on continuation, completion and progression make this concrete, growth that breaks those isn't growth. Where AI helps: selection fit, consistency, speed of decision making, and early identification of at-risk students in the first term. Where it doesn't help: manufacturing teaching capacity. Worth also considering modes that don't load the same infrastructure, modular and LLE-funded provision, online delivery, degree apprenticeships, rather than piling more bodies into the same lecture theatres.Alistair Sergeant

The AI and automation maturity gap

Most institutions are aiming for full, joined-up, intelligent automation. In reality, most are still at the early stages.

AI and Automation Maturity in Education, five stages from Manual to Optimising

Where does your institution actually sit?

Download the full breakdown: what each stage looks like in practice, and what it takes to move from one to the next.

Download the maturity model →

Build your own business case

You don't need a full process map to start. Identify two or three of your heaviest manual processes and put real numbers against them.

Admissions ROI Calculator

Put in your own numbers and get a realistic picture of the time and cost savings for your institution, no sales call required first.

Build your business case →

Use cases across the student journey

The specific processes institutions are automating today, each linked to the closest matching use case on niico.ai.

Case studies

Real institutions, real numbers.

Eastern Education Group

Eastern Education Group

400 interviews scheduled in 7 hours

400interviews scheduled, two weeks earlier than planned
7 hrsto do it, a process that used to take about a week
Read the case study →
Anglia Ruskin University

Anglia Ruskin University

Automation embedded across the institution

76,997transactions processed by live agents this year
12,487manual hours saved this year
Read the case study →

Resources

Everything referenced in the conversation.

Admissions 2030: The Intelligent Front Door to Higher Education

Co-authored by Ben Rogers and Lesley O'Keeffe. The full case for automation as the default, not the add-on, built on the systems you already run.

Read the whitepaper

AI and Automation in Education Guide

Your 101 guide to AI and automation in education: core concepts, key technologies, and first steps for successful adoption.

Read the guide

The EU AI Act Explained: A Practical Guide for Education Leaders

A practical whitepaper helping technology and transformation leaders navigate the EU AI Act and prepare their institution for responsible AI adoption.

Read the report

See what this could look like for your institution

Book a 30-minute walkthrough with the Niico team. No pitch deck, just a practical conversation about where automation could make a difference for you.

Book a walkthrough →